Effect of changing the distribution of data in a Finite Mixture of Semiparametric Generalized Linear Models

Usually, data analysis in a finite mixture of statistical models using some of these distributions is complex. One of the distributions which in practice used a lot, but it has a problem in a finite mixture of statistical models is the Poisson distribution. In this paper, due to non-equality of the mean and variance of Poisson distribution for the real data, finite mixture of generalized linear models for negative binomial distribution as alternative distribution has examined. Finally, as application of the proposal model in negative binomial distribution , the equating of Ph.D. test score between years of 92 and 93 in national organization for educational testing has considered. Also, result of simulation for data, has been shown the advantage of negative distribution to Poisson distribution, also the important variables has been determined.